Traditionally telephony networks have been synchronous in nature. They were initially constructed for the purpose of carrying narrowband voice signals, as wideband and broadband traffic, such as data and video, did not exist. Today, both network operators and equipment vendors face the challenge of handling wideband and broadband traffic with an infrastructure that is essentially inadequate because it was not designed for this purpose.
With the proliferation of the Internet and the integration of Internet traffic onto telephony networks, the networks are now dealing with large volumes of data traffic. As the Internet grows and more Internet service providers arise, telephony service providers and Internet service providers are investing more money in capital equipment in order to provide more bandwidth and paths for the Internet service traffic to flow.
One way Internet Protocol data could be routed in through a telephony network is over an ATM network. However, such a method of transmission of Internet Protocol data is inefficient due to cell header, packet-to-cell adaptation and encapsulation overhead.
Another method of routing Internet Protocol data is over a SONET network. Traditionally, if an Internet service provider point of presence would be connected into a SONET network, it would be through a device such as an add/drop multiplexer. If that Internet service provider point of presence was to be connected to another Internet service provider point of presence, a dedicated STS link within the SONET network would need to be established between the two points of presence. As more and more Internet service providers are added to such a network, the number of dedicated STS links dramatically increases, causing allocation of network bandwidth to become very inefficient.
Recently, the Internet Engineering Task Force approved a draft of a new standard known as Multiprotocol Label Switching (MPLS). MPLS attempts to merge layer 2 and layer 3 data switching. Under MPLS, an edge node converts regular packets into an MPLS format. The edge node also handles priority. Core nodes, however, do not need to look deeper then the assigned label to perform a switching function. This speeds up the switching process because the core nodes no longer need to look at a packet's source and destination address and priority. The draft MPLS standards are available to the public for download from the Internet at www.ietf.cnri.reston.va.us/lD.html. Currently available drafts include 1) draft-ietf-mpls-arch-04.txt; 2) draft-ietf-mpls-framework-02.txt, 3) draft-ietf-mpls-label-emcaps-03.txt; 4) draft-ietf-mpls-fr-03.txt; 5) draft-ietf-mpls-ldp-03.txt; 6) draft-ietf-mpls-ldp-03.txt; 7) draft-ietf-mpls-rsvp-ldp-tunnel-01.txt; and 8) draft-ietf-mpls-bgp4-mpls-02.txt. Each of these draft MPLS standards is hereby incorporated by reference.
MPLS offers the Internet service provider community the ability to offer different grades of service. It also provides the ability to create virtual private networks through the stacking of labels. For instance, one label can designate a particular company, while a sub-label can indicate a specific group of users within that company. MPLS also makes it easier for Internet service providers to perform traffic engineering by permitting Internet service providers to explicitly route certain labels to alleviate congestion.
However, the MPLS standards do not address integration of MPLS switching into a SONET network.